Working Across Functions

The managerial question: outcomes without org-chart authority

Modern companies run as matrices. A product manager ships features that require engineering hours she does not control. A finance partner approves spend that operations must execute. A marketing lead launches campaigns that legal must bless and sales must deliver. None of these people manage all the contributors on their project chart. Yet leadership still holds them accountable for outcomes.

This is lateral leadership: delivering results through influence, clarity, and tradeoffs rather than through direct reports. From Unit 1's Business Functions and How They Interact, you know that value creation breaks at handoffs: sales promises what product cannot build, finance models assumptions operations cannot hit, legal requirements arrive after engineering committed. Unit 4 adds that decision rights and incentives often sit in different functions than the work itself. The cross-functional leader's job is to align those pieces without pretending to be everyone's boss.

Failure looks familiar. Initiatives slip while stakeholders "support" them in meetings but prioritize other OKRs (objectives and key results, goal-setting targets used by many companies). Blockers appear late because compliance was informed, not engaged. Resource fights become personality conflicts because expected value was never quantified. This lesson gives you maps and scripts to reduce those failures: stakeholder analysis, RACI (Responsible, Accountable, Consulted, Informed, a roles matrix for tasks), and negotiated trades grounded in economics.

Influence without authority: credibility, relationship, and cost

Influence is the ability to change others' priorities or behavior when you cannot command them. A practical decomposition:

Influence ≈ (Credibility × Relationship) × (Clarity of ask / Cost to them)

Credibility is evidence that you understand their world and deliver on commitments. Credibility is function-specific. Engineering leaders trust peers who write clear requirements and protect focus time. Legal trusts partners who flag risks early with facts, not surprises at launch. Credibility erodes when you overpromise, hide bad news, or treat every function as a service desk.

Relationship is accumulated trust from small interactions: you showed up prepared, you shared credit, you escalated appropriately instead of politicking in Slack. Relationship is not friendship. It is predictability under stress.

Clarity of ask means one sentence on what you need, by when, and why it matters to shared outcomes. Vague asks ("we need support") force the other function to guess and often default to no.

Cost to them is hours, risk, political capital, or opportunity foregone. If your ask consumes six engineering weeks during a freeze, you must acknowledge the cost and offer a trade.

From Unit 5's Adapting a Message to Stakeholders, the same recommendation needs different emphasis by audience. Engineering hears scope and sequencing. Finance hears EV (expected value, probability-weighted outcome). Legal hears risk and precedent. One influence conversation rarely works as a copy-paste email.

Influence grows in deposits before withdrawals. Hit a small date for a gatekeeper before asking for a large exception. Share data before asking for headcount. This is the same trust logic as Unit 4's execution systems: reliability is a currency.

Stakeholder mapping: power, interest, and position

A stakeholder map is a structured view of who matters for an initiative. Two axes dominate MBA practice:

Power: ability to allocate resources, set priorities, or veto (formal or informal).
Interest: how much the outcome affects their metrics, risk profile, or team.

Plot stakeholders mentally or on a grid. High power, high interest: manage closely (weekly touch, co-design). High power, low interest: keep informed with concise updates (do not overload). Low power, high interest: consult for ground truth (customer success, support leads). Low power, low interest: monitor lightly.

Then classify position toward your initiative: champion, supporter, neutral, skeptic, blocker, gatekeeper. Position is not permanent. Trades and early wins move people.

Maps fail when they are vanity posters. A map earns its keep when it drives strategy per stakeholder: what message, what trade, what sequence. If engineering is a blocker risk because capacity is zero, "more meetings" is not a strategy. A strategy names a deprioritized alternative project, a dedicated product manager for requirements, or a phased scope cut with finance-backed ROI (return on investment, gain relative to cost).

Connect mapping to Unit 1's Customers, Employees, Owners, and Other Stakeholders. Internal functions are stakeholders with conflicting legitimate interests. Finance protects margin and cash. Engineering protects stability and craft. Sales protects quota. Your job is not to eliminate conflict but to make tradeoffs explicit so leadership can decide.

RACI: one throat to choke, many hands to help

Ambiguous ownership is the silent killer of cross-functional work. Everyone is "involved." No one is accountable. RACI assigns roles per task or deliverable:

The critical rule: one A. Multiple accountable parties means nobody accountable. "The committee decided" is how launches slip.

RACI works best at deliverable level, not vague verbs. Weak row: "Launch." Strong rows: "API specification signed," "Customer communication approved," "Go/no-go decision recorded."

Pair RACI with Unit 4's Decision Rights and Organizational Coordination. RACI documents who decides; decision rights documents what decisions exist. A go/no-go without a named A is a recurring failure mode from Why Good Strategies Fail in Execution.

Consulted roles are expensive. Overusing C slows teams. Informed roles are cheap but must not substitute for consultation when risk is high (legal, compliance, security).

Resource negotiation as quantified trade

Cross-functional fights are usually resource allocation disputes dressed as philosophy. Engineering says no. Product hears arrogance. Often engineering is optimizing a portfolio you have not seen.

Make trades discussable by quantifying expected value and capacity. Example frame: "Payments v2 unlocks modeled ARR (annual recurring revenue, subscription revenue normalized to a year) of $2.4M with seventy percent confidence; internal tools save $200k cost at ninety percent confidence. We propose 120 eng-weeks to payments and 40 now plus 40 next quarter to tools versus slipping payments."

Unit 2's Unit Economics and Contribution Logic supplies the language: payback, margin, CAC (customer acquisition cost, fully loaded cost to win a customer). Unit 3's Prioritizing Problems by Impact and Urgency supplies sequencing when both options are positive NPV (net present value, value of cash flows minus initial outlay, adjusted for time).

Good negotiation offers real options: phase scope, split teams, time-box spikes, swap quarter priorities with executive sign-off. Bad negotiation repeats urgency without acknowledging opportunity cost.

Document trades in writing after verbal agreement. Memory diverges under pressure. A one-page decision memo (Unit 5) with options, recommendation, and named approver prevents rework.

Building trust deposits and recovery after conflict

Cross-functional work spans months. You will miss a date or step on someone's toes. Trust deposits are small delivered commitments: requirements doc on day one, compliance workshop scheduled before build, finance model shared before asking for headcount.

When conflict happens, separate person from incentive. "You blocked us" triggers defense. "Your OKR weights enterprise; my initiative is SMB. Can we escalate a joint trade to the GM?" invites problem solving.

After a failed launch retrospective, publish what will change in RACI and cadence. Unit 4's Execution Systems and Operating Cadence applies at initiative scale: weekly decision forum, visible backlog, definition of done that includes legal and support readiness.

Recovery also means apologizing for unclear asks without over-apologizing for legitimate priorities. "I should have engaged you in week one; here is the revised timeline and what I am deprioritizing to fund it" rebuilds credibility faster than blame.

LedgerFlow is a fictional B2B (business-to-business, selling to companies) payments software company. Product manager Alex must launch Payments v2 in ninety days. Dependencies: Engineering (eight weeks build), Legal (contract review), Compliance (audit trail sign-off), Product Marketing (PMM, product marketing manager), Finance (pricing approval).

Part A: Stakeholder map

Check: Every dependency with veto or capacity power is listed ✓

Part B: Strategy per stakeholder

1. Jordan (Eng): Trade deprioritization of internal tools Q3 for dedicated PM on requirements day one; offer phased launch (core API week eight, reporting week ten) to reduce perceived risk.
2. Sam (Compliance): Co-own checklist workshop week two on data flows; weekly fifteen-minute checkpoint vs surprise review at week seven.
3. Rina (Legal): Reuse standard payment module contract language to cut review from three weeks to ten days; Alex supplies redlines, not blank pages.
4. Pat (Finance): Share unit economics memo before pricing workshop; tie list price to modeled payback under twelve months.
5. Lee (PMM): Align launch story to compliance milestones so external comms do not promise features before sign-off.

Part C: RACI excerpt

Check: Each row has exactly one A ✓

Part D: Managerial read

If Alex treats Jordan as a generic "resource," Jordan will protect the roadmap Alex cannot see. Quantified trade ($2.4M ARR vs internal tools savings) elevates discussion to leadership level. Compliance engaged early converts Sam from gatekeeper-at-the-end to co-owner of requirements, reducing late veto probability.

Board question: "What did we stop to ship Payments v2?" If answer is unclear, expect hidden backlog growth and burnout. Operator question: "Is day-85 go/no-go tied to compliance checklist completion, not only feature completeness?"

RidgePay is a fictional payments infrastructure firm. Q3 engineering capacity: 400 eng-weeks total. Requests: Payments v2 (120 eng-weeks), internal tools refresh (80), fraud model upgrade (60), tech debt (40). Sum of asks: 300 eng-weeks without tech debt; 340 with partial tech debt. Only 400 available but maintenance consumes 100 eng-weeks, leaving 300 net for projects.

Part A: Net capacity

Requested projects sum to 300 if fraud (60) is fully funded and tech debt deferred. Internal tools at 80 plus payments at 120 plus fraud at 60 = 260; remaining 40 allows partial tech debt.

Part B: EV framing (finance model)

Expected value dollars (rough): Payments 1.68M, Tools 0.18M, Fraud 0.825M.

Part C: Proposed trade

• Payments v2: 120 eng-weeks (approved)
• Fraud upgrade: 60 eng-weeks (approved; regulatory pressure)
• Internal tools: 40 eng-weeks now, 40 Q4 (defer half)
• Tech debt: 40 eng-weeks (minimum viable)

Check: 120 + 60 + 40 + 40 = 260; 260 + 40 maintenance buffer already removed from net 300 → 40 eng-weeks slack for overrun ✓

Alternative rejected: slip Payments to Q4 loses modeled Q3 enterprise pipeline tied to v2 features ($900k EV at 0.6 confidence = $540k).

Part D: Managerial read

CFO view: portfolio obeys net capacity identity; slack is explicit, not fantasy. Eng VP view: deferred tools Q4 is logged with owner, not silent debt. Compliance view: fraud work funded reduces tail risk. Communication memo (Unit 5) should state deprioritized work by name to prevent "shadow commitments."

You are a data product lead launching a fraud scoring model in eight weeks. You need:

• Data Science (DS): two weeks of modeling time
• Security: one week review; their backlog is six weeks
• Engineering: three weeks integration

The DS team is committed to a CEO pet project for the next sprint. Security will not slip other clients without executive cover.

Tasks:

1. Build a minimal stakeholder map (role, interest, power, position) for DS lead, Security lead, Eng manager, CEO, and you.
2. Propose one concrete trade per blocker (DS and Security).
3. Draft a three-sentence influence ask to the Eng director for integration weeks, including clarity and cost acknowledgment.
4. Assign RACI rows for "model validation signed" and "production deployment approved."

Solution

1. Stakeholder map (example)

2. Trades

• DS blocker: Phased launch. Week 1–2 ship rules engine with Security review of twenty hours; full ML model week 8 after CEO project slips one sprint with CEO memo citing $X monthly fraud loss EV. Trade: accept partial lift now vs full lift later.
• Security blocker: Offer dedicated analyst for checklist pre-work; narrow scope to integration touchpoints only (not full pen test) with written acceptance of residual risk from CRO (chief revenue officer, executive owning sales).

3. Three-sentence ask to Eng director

"We need three eng-weeks between July 8 and August 2 to wire the fraud rules engine into checkout; without it, modeled expected fraud loss is $420k over Q3. I know sprint 23 is committed; if we slip the internal admin refactor by one sprint, I will staff a QA rotation so your on-call load does not rise. Can you confirm the three weeks if the CRO signs the fraud EV memo by Friday?"

4. RACI

Check: Each task has one A ✓

A marketing leader wants Legal to approve ten custom customer contracts by month-end. Legal has two lawyers; average review is three days each; they already have twelve reviews queued (thirty-six lawyer-days needed; twenty-two lawyer-days available in the month assuming two lawyers, eleven days each after overhead).

Tasks:

1. Compute capacity gap in lawyer-days.
2. Propose two trades (scope, template, or executive) that close the gap without hiring.
3. Explain why RACI "A = Legal" for each contract may be insufficient without a prioritized queue owner.

Solution

1. Capacity gap

Available: 2 lawyers × 11 net days = 22 lawyer-days
Demand: (12 queued + 10 new) × 3 days = 66 lawyer-days
Gap: 66 − 22 = 44 lawyer-days short ✓

2. Trades (examples)

• Template trade: Move eight of ten to standard module (one day each) → saves sixteen days on those eight; remaining two custom at three days = six days; total new work eight days vs thirty days.
• Executive trade: CEO/CRO deprioritize six low-ACV (annual contract value, yearly revenue per customer contract) queued reviews to next month, freeing eighteen days for strategic ten.

Combined, demand drops toward capacity; exact mix needs leadership queue call.

3. RACI insufficiency

Legal can be accountable for review quality, but without a single queue owner (often commercial ops or sales ops) who sequences by EV and negotiates standard terms upstream, every sales rep escalates "urgent" deals. RACI tells who approves; it does not prioritize among exceeds-capacity work. Unit 4 decision rights must name who can say no to low-value custom deals.

Key takeaways

• Matrix roles require lateral leadership: credibility, relationship, clear asks, and acknowledged cost.
• Stakeholder maps must drive per-person strategy, not wall art.
• RACI enforces one accountable owner per deliverable; consult sparingly, inform generously.
• Resource fights become solvable when capacity and EV are on the table together.
• Trust deposits and written trades prevent cross-functional work from decaying into blame.

1. Draw a stakeholder map for your top cross-functional priority; label champions, gatekeepers, and blockers.
2. Identify one resource conflict and draft a quantified trade memo with two options and a recommendation.
3. Continue to Lesson 3: Developing Commercial Awareness.